Oceanographic drivers of legal-sized male Dungeness crab in the California Current System

We investigate environmental drivers of pre-season abundance of US West Coast legal-sized male Dungeness crab (Cancer magister), with the goal of developing an environmental index that can be used to forecast crab abundance in advance of the fishery. A conceptual life history approach is used to generate life-stage-specific and spatio-temporally-specific mechanistic hypotheses regarding oceanographic variables that influence survival at each life stage. Linear models are fit using the logarithms of pre-season abundance estimates of the coastal population of legal-sized male Dungeness crab as the dependent variable and environmental drivers from outputs developed using a regional oceanographic model for the California Current System as the independent environmental variables. Using different model selection methods, we show that the so-called best models differ substantially among model selection approaches, illustrating the need to carefully choose performance metrics for model selection. Since our goal was to forecast crab abundance, we selected the best model using a cross-validation metric that accounts for the time-series nature of the data. The resulting best models suggest that the mechanisms that drive preseason abundance differ among regions widely recognized for spatially and seasonally varying dominant physical processes. We found that the processes determining pre-season abundance of legal-sized male Dungeness crab could be identified with sufficient precision to enable a predictive skill, suggesting that the predictions may be useful for management purposes. Moreover, we found that transport (within and between regions), as well as temperature were likely drivers of pre-season abundance, highlighting that future studies should focus on multiple processes.